Concentration-Dependent Patterns of Leucine Incorporation by Coastal Picoplankton

Sampling site.

Surface water samples (1-m depth) were collected on 11 March and 6 May 2004 at the sampling station Helgoland Roads (54°11′N, 7°54′E; water depth, 8 m), 50 km offshore in the German Bay of the North Sea. The salinity was comparable at both sampling time points. In March the surface water temperature was 4.2°C and the phytoplankton biomass was low (<10 μg C liter⁻¹). The algal community was dominated by diatoms, in particular Thalassiosira nitzschioides. At the second sampling time point the water temperature was 8.3°C and the phytoplankton biomass was >200 μg C liter⁻¹. Although diatoms were still present, the phytoplankton community in May was dominated by colonies of the haptophyte Phaeocystis sp., which formed approximately 70% of the total algal biomass (Mursys Umweltreport, www.bsh.de).

Incubations with [³H]leucine.

Two sets of incubations were performed for the analysis of concentration-dependent leucine incorporation by the picoplankton assemblage. One set of samples, subsequently referred as bulk incorporation, was used to determine the rate of substrate incorporation by the total microbial community at different concentrations of [³H]leucine. The other set served to track the incorporation patterns of different populations by MAR-FISH. Both sets of samples were incubated with [³H]leucine (specific activity, 2.26 TBq/mmol; Amersham Bioscience, Freiburg, Germany) in a concentration series (0.1, 1, 10, and 100 nM) within 1 h after sample collection. For every treatment type, triplicate 10-ml subsamples plus one control, seawater fixed with freshly prepared buffered paraformaldehyde fixative (final concentration, 1%, vol/vol) (33), were incubated. The incubations were run for 4 h in the dark at ambient water temperatures. Subsequently, paraformaldehyde was added to the incubation mixtures to a final concentration of 1% (vol/vol).

After 1 h of fixation at room temperature the samples were frozen for transport and stored at −20°C. The samples for bulk incorporation were filtered onto cellulose mixed-ester filters (type GSWP; pore size, 0.2 μm; diameter, 25 mm; Millipore, Eschborn, Germany) and rinsed twice with ice-cold trichloroacetic acid and ethanol (20). The samples for MAR-FISH analysis were filtered onto polycarbonate filters (type GTTP; pore size, 0.2 μm; diameter, 25 mm; Millipore, Eschborn, Germany) and rinsed twice with sterile phosphate-buffered saline. All filters were stored at −20°C until further analysis.

Bulk measurements of substrate incorporation.

The cellulose filters were placed in 6-ml scintillation vials and dissolved with 0.5 ml of ethyl acetate. Subsequently, 5 ml of Lumasafe Plus scintillation cocktail (Lumac LSC BV, Groningen, The Netherlands) was added to each vial. The amount of radiolabeled substrate assimilated by microbial cells was measured by scintillation counting (Packard TriCarb 2900 TR; Perkin-Elmer, Wellesley, MA). Measurements were corrected for quench (external standard method) and by subtraction of counts from the prefixed controls.

Quantification of populations by FISH.

Samples for the determination of the in situ abundance of the different bacterial populations were immediately fixed after collection with paraformaldehyde (final concentration, 1%, vol/vol). The abundance of different microbial taxa was determined by FISH with horseradish peroxidase-labeled oligonucleotide probes and catalyzed reporter deposition (CARD-FISH) (31).

In addition to probe EUBI-III (most Bacteria) (9), four specific probes were selected that targeted substantial populations (>1% of total counts) which also scored positively for substrate incorporation: EURY806 (marine group II Euryarchaeota) (44), ROS537 (members of the Roseobacter clade of α-Proteobacteria) (11), SAR86-1245 (members of the SAR86 clade οf γ-Proteobacteria) (10), and CF6-1267 (members of DE cluster 2 of the Flavobacteria-Sphingobacteria group of the phylum Bacteroidetes) (23). The eubacterial antisense probe NON338 (3) was used as a negative control. All probes were purchased from Biomers.net (Ulm, Germany). Counterstaining of CARD-FISH preparations was done with 4′,6′-diamidino-2-phenylindole (DAPI; final concentration, 1 μg ml⁻¹). Evaluation of DAPI- and FISH-stained cells was carried out by epifluorescence microscopy (100× magnification, Axioplan; Carl Zeiss, Jena, Germany) as previously described (33), counting a minimum of 1,000 DAPI-stained cells per sample.

MAR-FISH analysis.

A recently developed MAR-FISH protocol was applied to study substrate incorporation by specific picoplankton groups (2). The key features of this new protocol are the combination of CARD-FISH staining and microautoradiography of microbes on membrane filters without prior cell transfer onto glass slides and evaluation by semiautomated image acquisition and analysis. Triplicate samples from every treatment type were analyzed by MAR-FISH with the different probes as described previously (2), and controls were evaluated with probe EUB I_III. Altogether, 108 MAR-FISH-stained preparations were evaluated.

The photochemicals for MAR (autoradiography emulsion type NTB-2, developer type Dektol, and fixer) were purchased from Eastman Kodak, Rochester, N.Y. Development of the exposed slides was performed according to the manufacturer's instructions (2 min of development, 10 s of rinsing with deionized particle-free water, 5 min exposure in fixer, washing for 5 min in deionized particle-free water). Different MAR exposure times were used for the evaluation of different treatment types: March samples, 72 h (0.1 nM), 48 h (1 and 10 nM), and 24 h (100 nM); May samples, 72 h (0.1 nM), 24 h (1 and 10 nM), and 12 h (100 nM). This strategy was chosen to obtain the maximal number of cells scored as having silver grains (subsequently referred to as leucine active) at different concentrations of [³H]leucine rather than to quantitatively compare incorporation under one exposure condition (e.g., by determination of grain size area) (7). Images of MAR-FISH preparations were captured with a digital camera (Orca I; Hamamatsu, Herrsching, Germany) mounted on a motorized epifluorescence microscope (Axioplan II Imaging; Carl Zeiss). Evaluation was performed as previously described (2) using the image analysis software KS400 (Carl Zeiss).

Hypotheses for statistical analysis and statistical tests.

The following formal hypotheses were tested. (i) The fractions of leucine-active Bacteria (as detected by probe EUB I-III) and leucine-active cells affiliated with the SAR86 and Roseobacter clades and with marine Euryarchaeota would be different between the two sampling times. (ii) The fractions of all leucine-active Bacteria and of members of the SAR86 and Roseobacter clades, of DE cluster 2 and of marine Euryarchaeota would be different at different concentrations of [³H]leucine. This was tested separately for each probe-defined cell population (dependent variable) by two-way linear analysis of variance of the sets of triplicate determinations, with concentrations and sampling time points as independent variables, and by post-hoc pairwise multiple comparisons using Bonferroni corrections. For DE cluster 2, which was present only in May, hypothesis ii was tested by one-way analysis of variance (concentration) and post-hoc comparisons (Bonferroni's method). Calculations were performed using the software SigmaStat (SPSS Inc., Chicago, Ill.).

Community contribution of the study populations.

Total cell numbers did not change substantially between the sampling dates: 0.8 (± 0.2) × 10⁶ cells ml⁻¹ in March and 1.2 (± 0.4) × 10⁶ cells ml⁻¹ in May. At both time points 84% of all DAPI-stained objects were detected by CARD-FISH (Table 1). Euryarchaeota constituted a constant fraction of total counts at both sampling dates. In May, the contribution of Roseobacter to all hybridized Bacteria was slightly higher. Members of the SAR86 clade formed a similar fraction of hybridized Bacteria and of total cell numbers at both months. Members of DE cluster 2 of the Flavobacteria-Sphingobacteria group, as targeted by probe CF6-1267, were too rare for accurate quantification in March, and they represented 5% of hybridized Bacteria in May.

Incorporation of labeled substrate.

Bulk incorporation rates of leucine were higher in May at [³H]leucine concentrations of >0.1 nM (Fig. 1A). In March the amount of incorporated leucine was similar at all concentrations, whereas in May it steeply increased with increasing substrate concentration.

A significantly higher proportion of bacterial cells incorporated tritiated leucine in May than in March (F = 72.9, P < 0.0001), in particular at the lower substrate concentrations (interaction of variables date and concentration, F = 9.5, P < 0.001) (Fig. 1B). The proportion of bacterial cells with visible incorporation of [³H]leucine did not significantly increase at concentrations of >10 nM in March and >1 nM in May (Bonferroni pairwise comparisons, P > 0.05). Negative controls from prefixed samples yielded similar percentages of false-positive MAR-active cells irrespective of sampling time point, incubation, and exposure conditions; approximately 1% of cells hybridized with probe EUB I-III (data not shown). In both months, the leucine incorporation rate per leucine-active bacterial cell was fivefold higher at 1 nM than at 0.1 nM added [³H]leucine. In March this parameter did not increase further with concentration, whereas in May it increased by a factor of 30 across the total range of concentrations (Fig. 1B).

The majority of members of the Roseobacter clade showed visible substrate incorporation at all substrate concentrations (Fig. 2A). A higher fraction of cells from this population incorporated leucine in May than in March, as was observed for the whole bacterial assemblage (Fig. 1B). However, in contrast to Bacteria, the proportion of leucine-active cells affiliated to Roseobacter was already high at the lowest concentration of added [³H]leucine and there was no (May) or only a little (March) increase in the fraction of Roseobacter cells with visibly incorporated [³H]leucine at increasing substrate concentration (Fig. 2A). Although members of the Roseobacter clade constituted only 8 and 10% of total hybridized cells in March and May, respectively (Table 1), these bacteria represented >30% (March) and almost 50% (May) of all leucine-active Bacteria at the lowest concentration of added [³H]leucine (Fig. 2B).

Members of DE cluster 2 of the Flavobacteria-Sphingobacteria group were only present in May. The number of cells with visible incorporation of [³H]leucine increased with substrate concentration (Fig. 3). The percentage of leucine-active cells approximately doubled between [³H]leucine concentrations of 1 and 10 nM, and their respective proportions were very similar at the lower two and at the higher two concentrations (Fig. 3). The fraction of cells from the SAR86 clade that incorporated the radiolabeled substrate also increased with increasing concentration (Fig. 3). In contrast to Bacteria and to the other populations studied, the proportion of leucine-active SAR86 cells at a given concentration of [³H]leucine was very similar at the two sampling time points. The only exception to this was at 1 nM added [³H]leucine, where the percentage of MAR-positive cells was significantly higher in May (Fig. 3) (Bonferroni pairwise comparisons, P > 0.05).

A high proportion of cells from the marine group II Euryarchaeota were able to incorporate [³H]leucine at all concentrations (Fig. 3). As observed for Bacteria, this fraction was significantly higher in May, when up to 70% of all Euryarchaeota were also MAR positive (Fig. 3). The percentage of cells incorporating leucine significantly increased at substrate concentrations >1 nM in March and >0.1 nM in May.

The proportion of cells with visible leucine incorporation was two- to fourfold higher in members of the Roseobacter clade than in hybridized Bacteria, and it was on average twice as high in Euryarchaeota (Fig. 4). This was particularly pronounced in March and at the lowest substrate concentrations. By contrast, the fraction of leucine-incorporating cells in bacteria affiliated with the SAR86 and DE cluster 2 was significantly below the community average at the three lower substrate concentrations. It matched the average fraction of all leucine-active Bacteria at a leucine concentration of 100 nM (Fig. 4).

Substrate incorporation by the whole microbial assemblage.

On average, approximately 30% of all bacterial cells in our samples were able to incorporate [³H]leucine. This has been suggested to be a typical value for aquatic systems (43). However, we also show that the fraction of cells that are “active” with respect to incorporation of this amino acid may be related to the concentration of the offered substrate (Fig. 1B), in particular if specific populations are investigated. If we had used only one concentration of [³H]leucine, we would have concluded that 10% (1 nM), 20% (10 nM), or 30% (100 nM) of bacteria from the SAR86 clade are “active” in March (Fig. 3).

At substrate concentrations of 1 to 10 nM the proportions of leucine-active prokaryotic cells (Fig. 1B) and the rate of leucine incorporation (Fig. 1A) were significantly higher during the bloom than before. This agrees with observations that blooms of Phaeocystis may stimulate microbial utilization of monomeric substrates (25). In the prebloom situation the incorporation rate per leucine-active microbial cell remained constant at concentrations of 1 nM added [³H]leucine and above, whereas it steeply increased over the whole range of offered substrate concentrations in May (Fig. 1B). The incorporation rates per leucine-active cell were very similar at the lower two concentrations (0.1 and 1 nM) in both months (Fig. 1B). By contrast, substantially more cells could incorporate the substrate even at low concentrations during the Phaeocystis bloom, and these “active” bacteria (and possibly also Euryarchaeota) were able to incorporate higher concentrations of substrate at significantly higher rates (Fig. 1B). This suggests a transition from an “oligotrophic” (high affinity) to a more “copiotrophic” (high velocity) substrate incorporation pattern, as previously observed during a spring bloom in the Chukchi Sea (47).

Substrate incorporation patterns of single picoplankton groups.

Incorporation of tritiated leucine was observed in all studied prokaryotic populations at all substrate concentrations (Fig. 2 and 3). We selected different MAR exposure times to obtain a maximum number of leucine-active cells. In our experimental setup a fixed exposure time would have resulted either in underestimation of leucine-active cells at low concentrations of [³H]leucine or in complete blackening of preparations at high concentrations. Potentially, our strategy might have diminished the differences between leucine-active cells at the various substrate concentrations. However, since we nevertheless detected significant differences in the incorporation patterns of the phylogenetic groups studied, the main hypothesis of our investigation could be readily addressed by our approach.

At least two distinct patterns of incorporation could be distinguished. For one, a large fraction of cells from the Roseobacter clade were involved in leucine incorporation in May irrespective of the substrate concentration (Fig. 2A). By contrast, a higher number of leucine-active cells were observed in bacteria affiliated to the SAR86 clade and DE cluster 2 at increasing concentrations of [³H]leucine (Fig. 3). Euryarchaeota (and Roseobacter during the prebloom situation) ranged between these extremes; i.e., the relationship between the proportion of leucine-active cells and substrate concentration was weak but significant (Fig. 2A and 3). Interestingly, none of the bacterial populations studied displayed the pattern that was characteristic of the bacterial assemblage as a whole (Fig. 1B, 2A, and 3). However, the three bacterial groups examined here only accounted for 25 to 50% of all leucine-active bacterial cells. Therefore, it cannot be excluded that other populations more closely resembled the “average” pattern of leucine incorporation.

Members of the Roseobacter clade are often found during natural or induced phytoplankton blooms (35, 38, 48). These bacteria are associated with, e.g., diatoms (17, 39) or dinoflagellates (36), and they colonize alga-derived detrital particles (38). In addition, bacteria from this clade also thrive under very low nutrient conditions (34). Our data suggest that members of the Roseobacter lineage in coastal North Sea waters are good competitors for leucine across a range of concentrations (Fig. 4) and that these bacteria may thus profit from substrate gradients as encountered during a Phaeocystis bloom situation.

Some genotypes within the Roseobacter lineage are free-living, whereas others preferentially attach to particles (12, 17). Therefore, it is possible that physiologically distinct subpopulations were responsible for [³H]leucine incorporation at different substrate concentrations in our samples. However, the constantly high percentage of leucine-active cells at all substrate concentrations (Fig. 2A) indicates that the Roseobacter population studied behaved as a single ecophysiological unit with respect to leucine assimilation.

SAR86 is a phylogenetic lineage of free-living bacteria in coastal and offshore waters (1, 10, 27). These bacteria have not been described from algal cultures (17, 35, 39), but they have been detected in high quantities during a coccolithophore bloom (15). In the German Bight, SAR86 bacteria grew more rapidly in late summer than during spring (30), and they did not incorporate glucose offered at a concentration of 10 nM (2). The percentage of leucine-assimilating cells from the SAR86 group in our samples was below the community average at all but the highest concentration (Fig. 4). This suggests that SAR86-related bacteria are competitive for leucine only if its supply is high. The similar incorporation patterns of members of the SAR86 clade during the prebloom and bloom situations (Fig. 3) moreover indicate that the activity of SAR86 may not be governed primarily by the availability of alga- derived monomeric substrates.

Marine DE cluster 2 is a monophyletic group of highly similar (>99%) 16S rRNA sequence types closely related to genera such as Geldibacter and Salgentibacter. It was first described from the Delaware estuary and from the Chukchi Sea (23). Members of this lineage occurred at considerable densities in the coastal North Sea (Table 1). While the three habitats differ in many respects, they also have some common characteristics, e.g., seasonally high primary production, influence of river discharge, and a high exchange between the sediment and the water column (23, 37, 40, 46).

Bacteria of DE cluster 2 were present only during the Phaeocystis bloom, whereas they were too rare for FISH quantification in the prebloom situation (Table 1). This suggests that their occurrence may be related to the development of the phytoplankton assemblage. Other members of the Flavobacteria-Sphingobacteria group have been found during natural (12) and artificial (17, 35, 38) phytoplankton blooms, including Phaeocystis-dominated blooms (42). The significantly higher numbers of leucine-active cells from DE cluster 2 at higher concentrations of added [³H]leucine (Fig. 3) might indicate that these bacteria require the rich conditions provided by the phytoplankton bloom, but also that leucine was not their preferred substrate. Other free-living marine representatives of the Flavobacteria-Sphingobacteria group are known to degrade polymers or to preferably incorporate high-molecular-weight dissolved organic matter (8, 22).

The planktonic marine group II Euryarchaeota are a common element of microbial assemblages in coastal waters during the growing season of the phytoplankton (28, 32), whereas Crenarchaeota in North Sea surface waters only formed larger populations during winter (32). In the upper layers of the central North Atlantic, 20 and 30% of Euryarchaeota cells assimilated aspartate and leucine, respectively, and these microbes also incorporated radiolabeled bicarbonate (18, 44). In our samples, the relative abundance of leucine-active Euryarchaeota was always higher than that of Bacteria and twice as high at the lowest substrate concentration (Fig. 4). Their activity moreover increased at higher substrate concentrations and during the phytoplankton bloom (Fig. 3). Thus, these microbes appear to be highly competitive for dissolved free amino acid across a wide range of concentrations.

Conclusions.

We show that several microbial populations that compete for the same substrate differed in their concentration-dependent patterns of leucine incorporation. Thus, their coexistence might be favored by microscale or short-term temporal substrate patchiness. In fact, our observations provide indirect evidence for the ecological relevance of habitat variability. A large fraction of the dissolved free amino acid turnover might be carried out by a few specialists (e.g., Roseobacter) at low substrate concentrations, whereas other taxa with lower uptake affinity (e.g., members of DE cluster 2) might participate in this process during phytoplankton bloom situations. Our results moreover agree with the hypothesis that some heterotrophic microbial populations in marine picoplankton may harbor several uptake systems that operate at different ranges of substrate concentration.